| 考虑时滞的铁水硅含量预报模型 |
| |
| 引用本文： | 王炜,余方超. 考虑时滞的铁水硅含量预报模型[J]. 山东冶金, 2006, 28(2): 41-43 |
| |
| 作者姓名： | 王炜 余方超 |
| |
| 作者单位： | 武汉科技大学,钢铁冶金及资源利用省部共建教育部重点实验室,湖北,武汉,430081;马鞍山钢铁股份有限公司,安徽,马鞍山,243000 |
| |
| 摘 要： | 根据神经网络的特点,建立了高炉硅预报神经网络模型和学习算法来预报高炉铁水硅含量。根据高炉冶炼的实际生产数据,选取风温、风量、透气性、料速、炉顶温度、焦炭负荷、喷煤量、上一炉铁水的硅含量作为输入,并对输入参数进行时效和时滞处理,采取附加动量项和自适应学习步长的措施缩短系统学习时间,提高了预报的准确率。应用表明,当允许绝对误差不大于0.1时,命中率为85.25%。
|
| 关 键 词： | 铁水 硅含量 预报模型 神经网络 时滞处理 |
| 文章编号： | 1004-4620（2006）02-0041-03 |
| 收稿时间： | 2006-01-10 |
| 修稿时间： | 2006-01-10 |
A Model for Predicting the Silicon Content in Hot Metal Based on Neural Network and Considered the Time Lag |
| |
| WANG Wei,YU Fang-chao. A Model for Predicting the Silicon Content in Hot Metal Based on Neural Network and Considered the Time Lag[J]. Shandong Metallurgy, 2006, 28(2): 41-43 |
| |
| Authors: | WANG Wei YU Fang-chao |
| |
| Affiliation: | 1 .Key Laboratory of Ferrous Metallurgy and Resources Utilization, Ministry of Education, Wuhan University of Science and Technology, Wuhan 430081, China; 2. Maanshan Iron and Steel Company Limited, Maanshan 243000, China |
| |
| Abstract: | A model for predicting the silicon content in hot metal based on neural network is introduced.According to the practice production data of blast furnace smelting,blast temperature,blast volume,permeability,lowering speed,top temperature,burden,pulverized coal injection and last silicon content are selected as inputs.The inputs are treated because of the time lag and the hitting rate of predicting the silicon content has been increased.To resolve the problem from local convergence and much time in learning in BP neural network,some methods such as the adding momentum and the variable learning speed are adopted to improve the neural network.It was indicated by the predicting results that the hitting rate is 82.25% when the absolute error is less than 0.1. |
| |
| Keywords: | hot metal silicon content predictive model neural network time lag treatment |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
